The scope was every fused deposition modeling (fused deposition modeling (FDM)) printer sold new to a US consumer as of 2026-09-06, from any brand, at any price. Price decided how deeply a model was captured, never whether it was listed. Resin, laser and industrial machines are out of scope but still appear in the coverage record with a reason, because an omission and an exclusion look identical to a reader unless you say which one it is.
Brands were enumerated first from a manifest, then a discovery pass went looking for what the manifest had missed: retailer catalogs at Micro Center, MatterHackers, B&H, Amazon, Newegg and Walmart; editorial roundups from All3DP, Tom's Hardware, PCMag, Wirecutter and TechRadar published within twelve months; the vendor lists inside the OrcaSlicer and Klipper source repositories; community discussion; and trade coverage. Discovery found 62 brand names that were not in the original manifest. Almost all turned out to be children's printers, industrial machines, defunct companies or brands with no US channel, but they are all recorded rather than dropped.
Every non-null fact in the underlying data carries a source, a retrieval date and an evidence tier. There are 629 sources behind this report.
| Tier | What it is | Count |
|---|---|---|
| T1 | evidence tier 1 | 327 |
| T2 | evidence tier 2 | 70 |
| T3 | T3 | 55 |
| T4 | evidence tier 4 | 176 |
| T5 | T5 | 1 |
Specifications and prices require tier 1. Reliability, noise, purge waste and support claims require at least one tier 2 plus one tier 4, or two tier 2. A manufacturer's own claim about how quiet or how reliable its machine is was never sufficient on its own, and marketing figures for speed and acceleration never moved a score by themselves.
Each configuration is scored 0 to 10 on eight judgments against written anchors. Each judgment carries a weight; the total is the sum of score times weight, divided by the full weight sum.
| Judgment | Weight | Capped for insufficient evidence |
|---|---|---|
| Just-works reliability and support | 25 | 4 |
| Filament magazine: what it holds and how well | 15 | 0 |
| Value within budget | 15 | 4 |
| Openness and longevity | 15 | 4 |
| Room to grow (chamber, magazine, harder materials later) | 5 | 1 |
| Materials now (PLA, PETG, ASA, ABS and ABS-GF) | 15 | 2 |
| Software and ecosystem UX | 5 | 4 |
| Practicalities for a shared home | 5 | 33 |
The rule in that last column is the most important one in this report. A judgment whose evidence does not meet the bar is capped at 3.0 out of 10, and the denominator is always the full weight sum. A gap always costs; it never helps.
That sounds obvious and it is not. The previous version of this study divided by the number of judgments it could score, which is arithmetically identical to filling each gap in at the machine's own average. The result was a ranking that correlated negatively with how well documented a machine was, and whose top pick was a machine the same study's own research said to avoid. The rule exists because of that failure, and a test in the build fails if anyone reintroduces it.
That rule has a second-order trap of its own, and this study fell into it once. The first scoring pass capped 23 of 67 configurations on reliability, including every Bambu Lab P1S, a machine three years on sale with an enormous published record. That cap was measuring this study's own reading, not the printers. A second research pass went and gathered the missing evidence under an explicit instruction to look for evidence rather than for good news, adding 165 sources and 85 defect records; both affected judgments were then re-scored from scratch. Scores moved in both directions, which is the only reason to trust the exercise: the P1S rose sharply, and the QIDI Q2 fell from about 7 to 4.0 once its documented requirement to run with the enclosure open came to light.
Pre-order, crowdfunded and announced-but-unshipped machines were excluded outright rather than scored and flagged. A machine that cannot be delivered cannot be the answer to "what should I buy right now", and chasing them consumed a great deal of effort in the previous version of this study for nothing. They appear once each in the timing chapter so the question "should I wait?" can still be answered honestly.
Third-party bolt-on multi-material unit (Multi Material Unit (MMU)) kits do not count as a multi-color path. Fitting a color system to a printer whose manufacturer never designed for one is a soldering-and-configuration project, which is the opposite of this buyer's stated primary criterion. Machines whose only color option is such a kit remain visible in the unranked section with the kit named, so the exclusion is legible rather than a silent gap. Whole-printer kits from an established vendor are a different thing and are scored normally.
The point of this chapter is that you can tell the difference between a machine this study rejected and a machine it never looked at. The previous version could not, and silently missed several brands.
Every brand encountered, whether in the manifest, in a retailer's catalog, in an editorial roundup, in a slicer's vendor list, or anywhere else, has a row in the coverage record with a status and at least one source. There are 78 such rows: 10 mainstream brands, 21 secondary and niche brands, 5 adjacent brands confirmed out of scope, and 42 rows for brands that discovery turned up and the manifest had not listed.
| Category | Count |
|---|---|
| Brands with a coverage row | 78 |
| Brands fully captured | 24 |
| Printers recorded in detail | 102 |
| Configurations priced | 128 |
| Configurations scored and ranked | 67 |
| Configurations visible but not ranked | 61 |
The 61 unranked configurations are not hidden. They appear in the tool with the reason they are set aside, which is almost always that the manufacturer sells no multi-color path for them, or that they cost more than the $1,600.00 stretch ceiling.
Nine sources of information could not be read directly and are declared rather than papered over: Adorama's catalog refused two URL variants; Reddit and YouTube refuse direct page fetches, so search results and a structured video interface were used instead; the listing pages at Kickstarter, Indiegogo and the Formnext exhibitor directory are JavaScript applications that returned nothing; and the Zortrax, Voxelab, Siboor and US Phrozen storefronts each refused repeatedly. Where an equivalent existed (a reseller page, a global storefront, a structured interface) it was used and labeled as such.
Sixteen prices captured in the first pass turned out to be wrong, and all sixteen are recorded in the conflicts log with both values. Two mistakes explain nearly all of them. A product page's default-selected variant is frequently not the base product, so a bundle price gets recorded against a bare machine. That is how the Bambu P2S Combo was first recorded at $780 when it is $749. And several manufacturers run a separate US storefront from their global one at different prices, which is how six Anycubic machines and the entire Phrozen line were mispriced.
Two of those corrections removed a candidate that had looked affordable. The UltiMaker S3 is $2,999, not the $1,095 first captured; that figure belonged to a different product in the same page's related-items carousel. The Rat Rig V-Core 4.1 needs a separate $768.56 electronics kit that its headline price does not include, putting its cheapest working build at $1,916.17. One correction went the other way: the Phrozen Arco is in stock in the US at $840, verified by adding it to a cart and requesting a live shipping quote to a Massachusetts postcode.